CT beam hardening correction method, CT beam hardening correction device, and storage medium

The present application is a U.S. National Stage entry of PCT Application no. PCT/CN2021/116208, filed Sep. 2, 2021, which claims priority to and the benefit of China patent application no. CN 202110718133.2, filed on Jun. 28, 2021, the contents of each of which are incorporated herein by reference in their entireties.

The present disclosure relates to the technical field of computerized tomography (CT) devices and, in particular, to a projection-based CT beam hardening correction method, a CT beam hardening correction device, and a storage medium.

In general, X-rays are produced in bremsstrahlung effects and contain broad spectral components. Since low energy X-rays have relatively high attenuation coefficients and are easily absorbed, low energy X-rays attenuate more when passing through objects, and X-ray spectra become “harder” after passing through objects. This means that the proportion of high energy X-rays detected by a detector increases, resulting in a smaller projection value. Because of the beam hardening effect, the attenuation coefficient along the X-ray path is not constant, and this will result in inhomogeneity even in a CT image of a uniform object.

To compensate for this effect, a beam filter may be used to reduce low energy X-rays and increase the average energy of the beam. However, the beam filter will reduce the dose and thus require a higher scanning tube power output, and it will increase the average energy of the X-ray beam and affect low contrast detectability. In addition, it still fails to meet the CT image quality requirements in most clinical scenarios.

Conventionally, in a beam hardening correction method for a CT device, a polynomial fitting correction algorithm is generally applied based on a water phantom. Because water is a main component of a human body, and most beam hardening correction algorithms need to determine polynomial factors for each system, most existing methods for determining a polynomial factor need to determine the polynomial factors in an iterative manner or based on image reconstruction. Since the iterative process and/or image reconstruction are usually inefficient, the entire beam hardening correction process is quite time consuming. Due to time costs, it is difficult or almost impossible to perform the entire correction process for each system.

Therefore, it is desirable to use an effective method to obtain these factors.

In view of this, the present disclosure proposes a CT beam hardening correction method, a CT beam hardening correction device, and a storage medium. One purpose thereof is to calculate a beam hardening factor by using (e.g. only) projection data of a scanned object without having to perform image reconstruction before calculating the beam hardening factor or iteration, thereby improving calculation efficiency, so that such correction method can be applied to each system in a tuning process of a CT device without adding extra time.

According to one aspect of embodiments of the present disclosure, a CT beam hardening correction method is provided, including: scanning a plurality of phantoms of different sizes by using a beam emitted by a ray source of a CT device, to obtain measured projection data of the plurality of phantoms; calculating estimated theoretical projection data of each phantom based on an initial estimated position of the phantom relative to the ray source by using a theoretical projection data calculation model; calculating an actual position of each phantom relative to the ray source based on the measured projection data of the phantom and the estimated theoretical projection data of the phantom; calculating actual theoretical projection data of each phantom by using the theoretical projection data calculation model according to the actual position of the phantom relative to the ray source; obtaining a beam hardening correction calculation model, where the beam hardening correction calculation model represents a relationship among expected projection data of a scanned object, measured projection data of the scanned object, and a beam hardening correction factor; using the actual theoretical projection data of each phantom as the expected projection data of the scanned object, using the measured projection data of each phantom as the measured projection data of the scanned object, and calculating the beam hardening correction factor according to the relationship by using the beam hardening correction calculation model; and using measured projection data of a measured object as the measured projection data of the scanned object and corrected projection data of the measured object as the expected projection data of the scanned object based on the calculated beam hardening correction factor, and calculating the corrected projection data of the measured object according to the relationship.

In this way, the beam hardening factor can be calculated by using only the projection data of the scanned phantom without performing image reconstruction before the beam hardening factor is calculated and a reconstruction process, thereby improving calculation efficiency so that such method can be applied to each system in a tuning process of a CT device without adding extra time.

In an example according to this embodiment, the beam hardening correction calculation model is represented by the following polynomial expression:

where i+j≤3; where Prepresents the expected projection data of the scanned object, Prepresents the measured projection data of the scanned object, B represents an inherent attenuation value of a wedge filter, and frepresents the beam hardening correction factor.

In this manner, the polynomial expression of the beam hardening correction model is preset, and then the beam hardening correction factor is obtained according to the actual theoretical projection data of the phantom and the measured projection data of the phantom. Compared with a conventional method in which a polynomial factor is obtained by using an image obtained by scanning a homogeneous phantom and the polynomial factor is solved in an iterative manner, the method in this application can avoid image reconstruction and iterative processes, and can be applied to each system in a tuning process of a CT device without adding extra time.

In an example of this embodiment, the calculating the beam hardening correction factor according to the relationship includes: substituting the actual theoretical projection data Pof each phantom as the expected projection data of the scanned object, and the measured projection data P of each phantom as the measured projection data of the scanned object into the polynomial expression of the beam hardening correction calculation model to obtain the following expression (1):

where i+j≤3; where if V=P·(P+B)·B, expression (1) is expressed as the following matrix calculation formula:

In this manner, calculation complexity of the beam hardening correction factor fcan be simplified and accuracy of determining the beam hardening correction factor fcan be improved.

In an example of this embodiment, the calculating corrected projection data of the measured object according to the relationship includes: substituting the measured projection data Pof the measured object as the measured projection data of the scanned object and the corrected projection data Pof the measured object as the expected projection data of the scanned object into the polynomial expression of the beam hardening correction calculation model to obtain the following expression (2):

In this manner, attenuation of a unit distance between a high-energy ray and a low-energy ray in a beam emitted by the ray source along an X-ray path is basically the same, so that an image reconstructed from projection data detected by a detector (array) has little beam hardening artifact.

In an example according to this embodiment, the ray source is a scanning tube, and the total amount N of the pixel data is equal to the quantity of the phantoms×the quantity of positions of the phantoms×the quantity of scanning positions of the scanning tube×the quantity of the detector pixels.

In this manner, the beam hardening correction method in this application can be made more robust, and can be adapted to measured objects of different sizes, for example, adults and children of different heights.

In an example according to this embodiment, the scanning tube performs scanning at a predetermined angle interval. For example, the predetermined angle interval may π/6.

In this manner, calculation time can be reduced.

In an example according to this embodiment, the measured projection data of the plurality of phantoms includes measured projection data obtained by scanning each phantom at a plurality of phantom positions of the phantom and a plurality of scanning spherical tube positions.

In this manner, the beam hardening correction method in this application can be more robust.

In an example of this embodiment, the theoretical projection data calculation model is represented as P=μl=μl+μl, where Prepresents the theoretical projection data of the phantom, μand μare respectively attenuation coefficients of a housing of the phantom and a uniform filler in the housing, lis a path length of a ray that is in the beam and that is angled at β relative to a center line of the beam in the housing, lis a path length of the ray in the filler, and land lare calculated by using the following formulas:

where D represents a distance between the ray source of the beam and the center of the phantom, βrepresents an angle between a line connecting the ray source of the beam and the center of the phantom and a center line of the beam, β represents an angle between the ray and the center line of the beam, rrepresents a radius of the housing of the phantom, rrepresents a radius of the filler of the phantom, and d represents a vertical distance between the center of the phantom and the ray.

In this manner, the theoretical projection data of the phantom can be determined by using a simple calculation process.

In an example according to this embodiment, the initial estimated position of each phantom relative to the ray source of the beam includes an initial estimated distance D′ between the ray source of the beam and the center of the phantom and an initial estimated angle β′ between the line connecting the ray source of the beam and the center of the phantom and the center line of the beam, and the calculating estimated theoretical projection data of each phantom based on an initial estimated position of the phantom relative to the ray source of the beam by using a theoretical projection data calculation model includes: presetting the initial estimated distance D′ and the initial estimated angle β′; and substituting the initial estimated distance D′ as the distance D between the ray source of the beam and the center of the phantom and the initial estimated angle β′ as the angle βbetween the line connecting the ray source of the beam and the center of the phantom and the center line of the beam into the theoretical projection data calculation model to calculate the estimated theoretical projection data.

In this manner, the initial estimated position of each phantom relative to the ray source of the beam is estimated according to experience, and the theoretical projection data is estimated according to the initial estimated position, thereby simplifying a calculation process.

In an example according to this embodiment, the actual position of each phantom relative to the ray source of the beam includes an actual distance D″ between the ray source of the beam and the center of the phantom and an actual angle β″ between the line connecting the ray source of the beam and the center of the phantom and the center line of the beam, and the calculating an actual position of each phantom relative to the ray source of the beam based on the measured projection data of the phantom and the estimated theoretical projection data of the phantom includes: determining the actual distance D″ and the actual angle β″ by minimizing the sum of the differences of squares of the measured projection data and the calculated estimated theoretical projection data by using a simplex multi-parameter optimization method.

In this manner, the initial estimated position of each phantom relative to the ray source of the beam can be corrected based on the measured projection data of the phantom, and therefore, the actual position of each phantom relative to the ray source of the beam can be accurately determined.

In an example of this embodiment, the calculating actual theoretical projection data of each phantom by using the theoretical projection data calculation model according to the actual position of the phantom relative to the ray source of the beam includes: substituting the determined actual distance D″ as the distance D between the ray source of the beam and the center of the phantom and the actual angle β″ as the angle βbetween the line connecting the ray source of the beam and the center of the phantom and the center line of the beam into the theoretical projection data calculation model to calculate the actual theoretical projection data of each phantom.

In this manner, the actual theoretical projection data of the phantom can be accurately determined by reducing calculation complexity.

According to another aspect of the embodiments of the present disclosure, a CT beam hardening correction device is provided, including: an estimated theoretical projection data calculation module, configured to calculate estimated theoretical projection data of each phantom based on an initial estimated position of the phantom relative to the ray source by using a theoretical projection data calculation model; a phantom actual position calculation module, configured to calculate an actual position of each phantom relative to the ray source of the beam based on measured projection data of each phantom obtained by scanning a plurality of phantoms of different sizes by using a beam emitted by a CT device and the estimated theoretical projection data of each phantom; an actual theoretical projection data calculation module, configured to calculate actual theoretical projection data of each phantom by using the theoretical projection data calculation model according to the actual position of the phantom relative to the ray source of the beam; a beam hardening correction calculation model obtaining module, configured to obtain a beam hardening correction calculation model, where the beam hardening correction calculation model represents a relationship among expected projection data of a scanned object, measured projection data of the scanned object, and a beam hardening correction factor; a beam hardening correction factor calculation module, configured to use the actual theoretical projection data of each phantom as the expected projection data of the scanned object, use the measured projection data of each phantom as the measured projection data of the scanned object, and calculate the beam hardening correction factor according to the relationship by using the beam hardening correction calculation model; and a measured projection data correction module for a measured object, configured to: use measured projection data of a measured object as the measured projection data of the scanned object and corrected projection data of the measured object as the expected projection data of the scanned object based on the calculated beam hardening correction factor, and calculate the corrected projection data of the measured object according to the relationship.

In this manner, the beam hardening factor can be calculated by using only the projection data of the scanned object without performing image reconstruction before the beam hardening factor is calculated, and no iterative process is required in the calculation process. Therefore, calculation efficiency can be improved, so that such correction method can be applied to each system in a tuning process of a CT device without adding extra time.

According to still another aspect of the embodiments of the present disclosure, a storage medium storing a program is provided. When a processor executes the program, the program causes the processor to perform the foregoing CT beam hardening correction method.

It can be learned from the foregoing solutions that, in the present disclosure, the beam hardening factor can be calculated by using only the projection data of the scanned object without performing image reconstruction before the beam hardening factor is calculated, and no iterative process is required in the calculation process. Therefore, calculation efficiency can be improved, so that such correction method can be applied to each system in a tuning process of a CT device without adding extra time.

Reference numerals are as follows:

To make the objectives, technical solutions, and advantages of the present disclosure clearer, the following further describes the present disclosure with reference to the embodiments.

illustrates a flowchart of a CT beam hardening correction method according to an embodiment of this disclosure. As shown in, the CT beam hardening correction method according to this embodiment of the present disclosure includes the following blocks:

Block S: Scan a plurality of phantoms of different sizes by using a beam emitted by a ray source of a CT device, to obtain measured projection data of the plurality of phantoms.

The ray source may include a scanning tube, and the scanning tube may scan the phantom at a predetermined angle interval a (as shown in), for example, an angle interval π/6; and may scan each phantom relative to the ray source at different positions of the CT device. For example, the phantom may be scanned in a rotation center, and the phantom may be eccentrically scanned. Therefore, the measured projection data of the plurality of phantoms includes measured projection data obtained by scanning each phantom at a plurality of phantom positions of the phantom and a plurality of scanning tube positions.

Block S: Calculate estimated theoretical projection data of each phantom based on an initial estimated position of the phantom relative to the ray source by using a theoretical projection data calculation model (in, the ray source is an X-ray scanning tube, and the scanning tube is represented by a focal point thereof).

The theoretical projection data calculation model is expressed as the following formula:

Prepresents the theoretical projection data of the phantom, μand μare respectively attenuation coefficients of a housing of the phantom and a uniform filler in the housing, lis a path length of a ray that is in the beam and that is angled at β relative to a center line of the beam in the housing, lis a path length of the ray in the filler, and land lare calculated by using the following formulas according to the positional relationship among the phantom, the ray source, and the beam shown inand:

where

D is the distance between the ray source of the beam and the center of the phantom, βrepresents the angle between the line connecting the ray source of the beam and the center of the phantom and the center line of the beam, β represents the angle between the ray and the center line of the beam, rrepresents the radius of the housing of the phantom, rrepresents the radius of the filler of the phantom, and d is the vertical distance between the center of the phantom and the ray. Herein, the phantom may be a water phantom for simulating a human body. The housing of the phantom is made of PMMA materials, and the filler in the housing is water.

In this specification, the term “projection data” is used to indicate the attenuation amount.